Last Updated - 5th Feb 2025
Introduction
At AudioStack, ensuring our systems are developed, deployed, and used safely is our priority. Over the past six years, our AI audio production services have supported a wide range of applications—from consumer healthcare and news podcasts to video game character animation, sales support, language localization in entertainment, and dynamic voiceovers for social media and advertising. These technologies have not only transformed creative processes but have also enabled new business opportunities. We fundamentally believe in the transformative power of synthetic audio while recognizing that, as with any new technology, risks exist. Managing these risks through robust security and ethical protocols is central to our mission.
FAQ
Q: What services do you provide?
AudioStack (and its partner platform, Aflorithmic) provides the world’s first and only infrastructure for building synthetic audio. Users can create audio from text using the best AI voice providers available as well as proprietary models. Once speech is generated, it is transformed into audio using professionally designed sound templates and enhanced to a quality that rivals studio productions. This process brings automation and scalability to an industry that has changed little over the past century. We are committed to ethical, fair, and transparent AI, in accordance with the UK’s and European Union’s Ethics Guidelines for Trustworthy Artificial Intelligence—and we will adhere to emerging laws such as the AI Act.
Q: What are your protocols for Ethical AI?
We ensure that our development, deployment, and usage of AI systems meet the highest safety standards. By following established ethics guidelines and future regulations, we safeguard against misuse while promoting fairness and transparency in every aspect of our technology.
Q: What is your policy on Voice Cloning?
Consent Required: We will never use the voice of a private person or an actor without their explicit permission, captured through an audio recording.
Historical Figures: For historical figures, written consent from the estate is mandatory.
Open-Source Exception: Open-source technologies, such as vocoders (e.g., Stephen Hawking), are excluded from these restrictions.
Legal and Privacy Agreements: Voice cloning begins only after Aflorithmic (and AudioStack) and the respective voice actor or company have agreed on legal terms and privacy details.
No Self-Serve Cloning: Self-serve voice cloning is not permitted without direct communication with the individual concerned.
Vetted Providers: All AI voice providers are individually vetted and have signed agreements affirming their rights to commercially use the speech models they supply.
Secure Access: Access to cloned voices is restricted to authorized organizations through rigid authentication methods and is additionally secured by unique API keys.
Q: How do you handle Sound and Music?
Sound design templates and sound effects are crafted using either proprietary sound elements or those defined as creative commons. Any sound files uploaded by users are governed by our terms and conditions. Content identified as fraudulent, racist, sexual, or otherwise suspicious will result in the suspension of the associated account.
Q: How do you safeguard against misuse by Bad Actors?
Continuous Monitoring: We monitor nearly all created audio content daily.
Strict Account Controls: Any detection of fraudulent, racist, sexual, or otherwise suspicious content leads to immediate account suspension.
Signup Restrictions: Temporary email addresses are disallowed, and signups are limited to one per IP address.
Transparency: Free tier accounts include a watermark (“Created with AudioStack”) to clearly identify the source of the audio content.
Dedicated Support: All paid accounts are supported by a dedicated technical account manager who actively monitors and assesses any unlawful usage on an individual basis.
Q: What limitations do you acknowledge with synthetic audio technology?
While synthetic audio technology offers vast possibilities, it is not immune to misuse. We accept that any technology can be exploited and consider it our responsibility to limit abuse while educating the public about synthetic audio’s capabilities and risks. Consequently, any account generating fraudulent, racist, sexual, or otherwise suspicious content will be suspended.
Q: What is the Content Authenticity Initiative?
AudioStack (and Aflorithmic) is a proud member of the Content Authenticity Initiative (CAI), first announced by Adobe in 2019. This group comprises hundreds of creators, technologists, journalists, activists, and leaders committed to combating misinformation and enhancing content authenticity. With an open, extensible approach to media transparency, CAI enables better evaluation of content provenance by collaborating with representatives from software, publishing, social media companies, human rights organizations, photojournalism, and academia. For further details, please visit contentauthenticity.org.